In various examples, the present disclosure relates to using temporal filters for automated real-time classification. The technology described herein improves the performance of a multiclass classifier that may be used to classify a temporal sequence of input signals—such as input signals representative of video frames. A performance improvement may be achieved, at least in part, by applying a temporal filter to an output of the multiclass classifier. For example, the temporal filter may leverage classifications associated with preceding input signals to improve the final classification given to a subsequent signal. In some embodiments, the temporal filter may also use data from a confusion matrix to correct for the probable occurrence of certain types of classification errors. The temporal filter may be a linear filter, a nonlinear filter, an adaptive filter, and/or a statistical filter.
Legal claims defining the scope of protection, as filed with the USPTO.
2. The processor of claim 1, wherein the filter weights at least one classification output of the one or more classification outputs based at least on a recency of at least one classification output in the temporal series of classification outputs.
3. The processor of claim 1, wherein the filter weights at least one classification output of the classification outputs based at least on one or more classes being assigned to one or more second input signals of the input signals using at least one second filtered classification output of the temporal series of classification outputs.
4. The processor of claim 1, wherein the filter weights at least one classification output of the classification outputs based at least on detecting, using the temporal series of classification outputs, a classification state change.
5. The processor of claim 1, wherein the filter is based at least on one or more confusion factors corresponding to the one or more neural networks.
6. The processor of claim 1, wherein the filter includes one or more of a linear filter, a non-linear filter, an adaptive filter, or a statistical filter.
7. The processor of claim 1, wherein the temporal sequence of input signals is generated using one or more sensors of a machine, and the one or more circuits are further to perform one or more operations for the machine based at least on the classifying.
10. The system of claim 9, wherein the at least one classification output is filtered based at least on a recency of the at least one classification output in the temporal series of classification outputs.
11. The system of claim 9, wherein the at least one classification output is filtered based at least on one or more classes being assigned to one or more second input signals of the plurality of input signals using a second filtered subset of the temporal series of classification outputs.
12. The system of claim 9, wherein the at least one classification output is filtered based at least on detecting, using the temporal series of classification outputs, a classification state change.
13. The system of claim 9, wherein the at least one classification output is filtered based at least on one or more confusion factors corresponding to the one or more neural networks.
14. The system of claim 9, wherein the temporal sequence of input signals is generated using one or more sensors of a machine, and the operations further include performing one or more operations for the machine based at least on the classifying of the one or more input signals.
17. The system of claim 16, wherein the confusion factor corresponds to a probability the one or more input signals correspond to a second class of the plurality of classes.
18. The system of claim 16, wherein the adjusting the first confidence score using the confusion factor includes normalizing the first confidence score based at least on the confusion factor.
19. The system of claim 16, wherein the generating the second confidence score corresponding to the first class is based at least on filtering the first confidence score adjusted using the confusion factor.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 12, 2023
August 27, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.